DeepMind shows Gemini 4 Argon planning Spot's trash-bin task through a named skill
DeepMind reports that Gemini 4 Argon scores 87% on the ERQA embodied reasoning benchmark. Its Spot and robot-arm demos run through a named skill and a set of action primitives, so they test planning over a control stack that already exists.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+45
- Incentives70
- Confidence65
buildOne report1 publisher Apptronik's Apollo made seven straight throws from a simulation Gemini 4 Argon built out of video
Google says Gemini 4 Argon turned a video of a person throwing into a calibrated MuJoCo simulation, cutting setup from weeks to hours in an internal workflow. The speedup is Google's own measurement, as relayed by dev.to.
Reality
- Evidence30
- Adoption
- Not scored
- Hype gap+25
- Incentives70
- Confidence55
Safeworld raised more than $12 million, led by Shine Capital and a16z Speedrun, to test generative-AI robots against simulated people. The pitch is that makers of robots nobody can prove safe on paper will pay an outside tester to vouch for them.
Perspective Coverage
4 publishers
- Builder
- Builder 36%
- Operator
- Operator 34%
- Investor
- Investor 30%
Reality
- Evidence55
- Adoption15
- Hype gap+35
- Incentives75
- Confidence60
buildOne report1 publisher MJWarp scales one SO-101 arm to 2,048 worlds by changing what the thread index means
NVIDIA's walkthrough moves a MuJoCo follower arm up to 2,048 parallel GPU environments, where the kernel body barely changes and most of the work is model compatibility and an opt-in determinism mode.
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+5
- Incentives78
- Confidence60
Vsim's on-robot simulator checks 20,000 possible next seconds while Freddo moves
A Cambridge start-up says it trained walking and grasping in minutes. Its co-founder and Nvidia's robotics product director both put the current limit at fine dexterity and multi-step tasks like fill and pour.
Reality
- Evidence45
- Adoption15
- Hype gap+20
- Incentives70
- Confidence55
buildOne report1 publisher One GPU and a bash retry queue trained 14 skills for a bipedal robot duck
MicroDuck's 14 policies came out of roughly five days of serial single-GPU training in mjlab, with nine separate reward designs behind them and a VRAM ceiling that cut the env count to 2048 on rough terrain.
Reality
- Evidence45
- Adoption20
- Hype gap+15
- Incentives25
- Confidence55